Uploaded image for project: 'Horizon Stream'
  1. Horizon Stream
  2. HS-22

Provisiond in Horizon Stream



    • Epic
    • Status: To Do (View Workflow)
    • Major
    • Resolution: Unresolved
    • None
    • None
    • Provisiond in Horizon Stream


      With deploying a monitoring solution you want to start building monitoring coverage across your IT infrastructure. The following tasks create monitoring coverage:

      • running regular closed-box (black-box) tests against network services to measure availability, e.g. ICMP (ping), web requests, SMTP, IMAP, etc.
      • Collecting insight performance metrics provided by agents running on a system or within an application, e.g. SNMP, JMX, WS-MAN, Prometheus-Agent, or via streaming telemetry
      • Collection of network communication metrics through Network Flows or the BGP Monitoring Protocol
      • Collecting logs and error messages sent from devices to a monitoring system, e.g. SNMP Traps, Syslog

      The benefit of having good monitoring coverage can be compared with good test coverage in software, you gain confidence during maintenance and refactoring or rebuilding parts of your IT infrastructure as well as having measures to improve performance and increase availability and resilience. (You can't improve what you do not measure)

      Problem statement

      Network operators deploying a network monitoring solution for the first time don't know exactly in detail what they need to monitor and how they want to monitor it. The Provisioning system is a tool that tries to help people make this decision.

      How does it work in a nutshell?

      If a user creates a Node in a requisition, you can see it as a "seed" for the Provisioning process he tries to enrich it with the following information:

      • Running service detectors to identify network services or SNMP agents, e.g. ICMP, HTTP, SNMP, SMTP, IMAP, etc
      • In case the device has an SNMP agent running that can communicate with it, we enrich the Node with standard MIB-II system information, additional IP interfaces, and physical network interfaces. To give the user some control we have policies in place which help users to decide.

      How do we discover network devices?

      If we don't know "What to monitor" we have a "Discovery Daemon" which can scan IP network ranges with ICMP. Every reachable IP address is considered as a Node and fed into the Provisioning system (specifically a "Requisition") for further investigation.




            Unassigned Unassigned
            anaseef Arthur Naseef
            0 Vote for this issue
            2 Start watching this issue